Sundar Pichai unveils Gemma: Google’s cutting-edge Open AI models for developers

Google has introduced Gemma, the latest advancement in its AI technology, marking a significant step forward in open AI models. Google said that its name is derived from the Latin word ‘gemma,’ which translates to precious stone.

Sundar Pichai, CEO of Google and Alphabet, announced the launch of Gemma on the X platform, describing it as a family of lightweight, state-of-the-art open models designed for responsible AI development. 

Gemma is presented in two versions, Gemma 2B and Gemma 7B, each equipped with pre-trained and instruction-tuned models. 

Developed with common foundational elements and infrastructure components from the Gemini models, Gemma aims to provide developers and researchers with essential tools for AI development. 

Gemma integrates with popular tools like Colab and Kaggle notebooks and supports frameworks such as JAX, PyTorch, Keras 3.0 and hugging face transformers. 

Google emphasizes Gemma’s versatility, operating seamlessly across various platforms, including laptops, workstations and Google Cloud infrastructure. 

Google asserts that Gemma outperforms larger models across benchmarks while adhering to stringent safety and responsibility standards. 

In collaboration with Nvidia, Gemma is optimized for efficiency, offering developers the capability to build generative AI applications for tasks like text generation, summarization, and Q&A within Vertex AI and GKE environments.